WaifuGemma4 26B A4b V1 — Hardware Requirements & GPU Compatibility
ChatVisionRoleplay
WaifuGemma4 26B A4b V1 is a 25.8B-parameter open language model from hiwaifu-research in the Gemma 4 family. It supports a context window of up to 262,144 tokens. At BF16 it needs about 52.26 GB of VRAM — see which GPUs and Macs can run it below.
WaifuGemma4 26B A4b V1 runs on a 64 GB+ Mac or a workstation GPU
Best local pick → Mac Pro M2 Ultra (64 GB) · BF16 · 52.6 GB · ~46 tok/s
Macs from 64 GB (BF16 · ~46 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run WaifuGemma4 26B A4b V1
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve hiwaifu-research/WaifuGemma4-26b-a4b-v1 --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on A100 80GB (27 GB headroom) · BF16
- Generation speed
- ~120 tok/s
- generation speed
- Cost per 1M output tokens
- $2.70
- per 1M output tokens
How Much VRAM Does WaifuGemma4 26B A4b V1 Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 52.3 GB | 53.2 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 52.6 GB |
Selected BF16
Brain floating point 1652.6GBat 16K context
- Weights
- 51.6 GB
- KV cache
- 0.7 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 28.6 GB more
Or .
Which Hardware Can Run WaifuGemma4 26B A4b V1?
BF16 · 52.3 GBWaifuGemma4 26B A4b V1 (BF16) requires 52.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 68+ GB is recommended. Using the full 262K context window can add up to 8.6 GB, bringing total usage to 60.9 GB. No consumer GPU has enough memory. 52 devices with unified memory can run WaifuGemma4 26B A4b V1, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Pro M2 Ultra (64 GB).
64–80 GB
fits · 11.4 GB free · up to 256K ctx
3 Mac classes · 1 device family
M2 Ultra*64 GB~46Simulate output on Mac Pro M2 Ultra (64 GB)M5 Max64 GB~37Simulate output on Mac Studio M5 Max (64 GB)Orin~15Simulate output on NVIDIA Jetson AGX Orin 64GB
By VRAM tier · typical GPUs and Macs at BF16
Alternatives to WaifuGemma4 26B A4b V1
Gemma 4 family · 35 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.8 27BAlibaba | 122.0 | 17.9 GB | ~38 | same size |
| Qwen3.6 27BAlibaba | 111.2 | 17.9 GB | ~38 | same size |
| Qwen3.5 27BAlibaba | 103.8 | 17.9 GB | ~38 | same size |
| Magistral Small 2509Mistral AI | 85.6 | 18.1 GB | ~43 | same size |
| Gemma 3 27B ITGoogle | 79.4 | 21.7 GB | ~36 | same size |
| LFM2.5 24B A2BLiquid AI | — | 15 GB | ~221 | same size |
Specifications
- Publisher
- hiwaifu-research
- Family
- Gemma 4
- Parameters
- 25.8B
- Active parameters
- 4.4B per token
- Architecture
- Gemma4ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 262,144
- Release Date
- 2026-09-18
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does WaifuGemma4 26B A4b V1 need?
WaifuGemma4 26B A4b V1 requires 52.3 GB of VRAM at BF16. Full 262K context adds up to 8.6 GB (60.9 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 25.8B × 16 bits ÷ 8 = 51.6 GB
KV Cache + Overhead ≈ 0.7 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 9.3 GB (at full 262K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run WaifuGemma4 26B A4b V1?
No — WaifuGemma4 26B A4b V1 requires at least 52.3 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
Can I run WaifuGemma4 26B A4b V1 on a Mac?
Yes — Mac Pro M2 Ultra (64 GB) and 41 other Macs can run WaifuGemma4 26B A4b V1. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 52.3 GB of usable unified memory (RAM minus macOS overhead).
Can I run WaifuGemma4 26B A4b V1 locally?
Yes — WaifuGemma4 26B A4b V1 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is WaifuGemma4 26B A4b V1?
At BF16, WaifuGemma4 26B A4b V1 can reach ~127 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 8.9 GB active ÷ (8000 × 0.65) = 1.71 ms, plus 30 layers × 0.055 ms = 1.65 ms, so 1000 ÷ 3.36 ms = ~298 tok/s
Estimated speed at BF16 (52.3 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of WaifuGemma4 26B A4b V1?
At BF16, the download is about 51.61 GB.
Which GPUs can run WaifuGemma4 26B A4b V1?
No single consumer GPU has enough VRAM to run WaifuGemma4 26B A4b V1 at BF16 (52.3 GB). Multi-GPU or professional hardware is required.
Which devices can run WaifuGemma4 26B A4b V1?
52 devices with unified memory can run WaifuGemma4 26B A4b V1 at BF16 (52.3 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.